CFD, IEC STANDARDS AND TESTING LABORATORIES: JOINING THE PIECES

FOR HIGHER QUALITY HV EQUIPMENT.

Author name: Affiliation:

Sergio Feitoza Costa Cognitor Consultancy, Research and Training Ltd.

Summary: Electricity users are each time more exigent, on the need that the equipment they
use has a test certificate or report. Tests are expensive and usually the manufacturer, before
arriving to an approved design, need to repeat test some times. Simulation techniques which
include CFD ("computational fluid dynamics") are useful as a virtual laboratory in the
equipment design stage. Better simulations means less testing costs. IEC Standards are used
all over the World and are the basis for national standards. Within certain limits, simulation
techniques can be used to extrapolate the results of an already done laboratory test to other
similar, but not equal, equipment. This paper show examples of these possibilities and
present, for discussion, an objective strategy on "To establish some rules to enable simulation
techniques to be used as an auxiliary tool for the verification of some testing requirements in
IEC standards"

Keywords: CFD, Equipment, Testing, Arcs, Temperatures, Short-circuit, Ampacities, Switchgear

1. INTRODUCTION
To better understand some ideas which are proposed in the text to follow it is useful to inform
that the authors basic experience is on high power testing, IEC work management and
electrical equipment and substations design.
High power and high voltage testing laboratories are within the main tools a country has to
achieve a high electricity quality. Electricity users are each time more exigent, on the need
that the equipment they use has a test certificate or report. Such documents are used to
demonstrate that the product was type tested according to a recognized technical standard like
the IEC -International Electrotechnical Commission ones.
Type tests are done in qualified testing laboratories and usually the manufacturer, before
arriving to an approved design, need to repeat tests 2 or 3 times. In 2007 there are relatively

(*)
Cognitor - Consultancy, Research and Training Ltd - R.Babau, 103/202 - Rio de Janeiro Brazil - 21931-230 - www.cognitor.com.br
few laboratories all over the World. The clear trend is to have each time less laboratories
owned by the few companies which understood that this activity can be an interesting
business from the financial point of view. All over the World, we can count these ones
without needing to use all the fingers of just one hand.
Laboratory tests, specially the high power ones, are expensive because the laboratories
installations require investments in the order of magnitude of US$ 50 million. Many of the
existing laboratories are company dedicated ones and do not focus on selling services to
have a high occupancy to optimize the return on the investment done. This create situations
where there is a long waiting list and a manufacturer shall wait one year to have its equipment
tested and the test installation is occupied just 1/3 of the year hours.
We can give an idea on testing costs using metal-enclosed switchgear used on 15 to 36.2 kV
systems. To do only a part of them, comprising the "Temperature Rise Tests", the Short-
Time Current Withstand Tests" and " Pressure Relief during Internal Arc Tests", we arrive to
more than U$D15000,00. Adding transportation costs and few test repetitions we go well
beyond U$D20000, 00. This is the value if all the tests are successful in the first trial (a rare
fact on the first of a series). The uncertainties about succeeding put many manufacturers out
of the market.
Supposing that a certain type of equipment design was successfully tested it is a common fact
that the equipment purchaser request the manufacturer to test another design if the
geometrical is similar but different from the tested one. This occurs mainly when there are not
written calculation methods available to do the extrapolations.
Simulation techniques and tools, including the CFD ("computational fluid dynamics") ones,
used by experienced designers, are useful as a virtual testing laboratory. Temperature rise and
arcing energy pressure relief are good examples of possible test simulations (Figures 1 and 2).

They consist in doing a 3D simplified model of the equipment and to associate to the model
the components materials and the boundary conditions like temperature, wind speed, power
injection and others. The key point is that it is possible to visualize temperatures and air flows
which would need hundreds of thermocouples or sensors in a real laboratory test.

Figure 1 Internal arc simulation Figure 2 Temperature rise simulation

The example of the internal arc test in metal-enclosed switchgear is interesting to explain
the possibilities of such simulations and it will be described in more details in Section 2. Most

2
of the manufacturers have arrived to approved designs by the trial and repeat method
because some needed calculations like where the hot particles will arrive was practically
impossible to do in the past. Nowadays, with simulations techniques, it is possible at least to
know the effect of each parameter involved reducing the number of trials before succeeding.

Our objective, in this case, is to provide guidelines on how to enable that simulation tools are
used to extrapolate the results of an already done laboratory test to other in a similar, but not
equal, equipment (for example the tested equipment have dimensions 1,0 x 1,2 x 2,0 m and
the untested one is 1,0 x 1,3 x 2,1 m )

IEC Standards are used all over the World and are the basis for national standards. IEC
standards are the possible compromise which can be achieved at a certain moment by
manufacturers, users and laboratories. It is impossible to attend simultaneously all the
interests and sometimes some small words and sentences which are clear to the IEC Working
Group participants elaborating the standard are not clear or even perceived by the big number
of the standard users all over the World. This is not good but sometimes is the only possible
way to enable that, in a next revision, the subject will be included in the standard in details.
The problem is when the next revision of the standard occurs more than 10 years after. Some
examples are showed in Section 2.

If it is considered useful to have some link between simulation techniques and testing, the
proper way to disseminate the idea and to trace guidelines is through IEC standards. A
strategy for this will be proposed

So, the objectives of this paper are

a) To show examples of simulation techniques applications that could be used as an auxiliary
tool to support some IEC standards application.
b) To present a suggestion for a strategy, to be discussed within Cigr, on how "To establish
some rules to enable simulation techniques to be used as an auxiliary tool for the verification
of some testing requirements in IEC standards"

In no part of this text the idea is to propose to replace tests by simulations. The proposal is to
start to consider and study how some simulation techniques can be used in parallel to type
tests to make the life easier.

3. REFERENCE SITUATIONS TO FACILITATE THE UNDERSTANDING

To facilitate the understanding of the proposal I describe in 3.1 and 3.2 two specific situations
related to high power testing on high voltage equipment. For each one there is a small
description of what can be simulated, the possibility of applying simulation techniques on
extrapolating testing results and also, using the opportunity, some relevant points which are
not quite clear for many standard users but can be better explored through simulations.

3.1) INTERNAL ARC TEST ON METAL-ENCLOSED SWITCHGEAR

ACCORDING TO IEC 62271-200 (High voltage switchgear and controlgear - Part 200: AC
metal-enclosed switchgear and controlgear for rated voltages above 1 kV and up to and
including 52 kV)

3
IEC 62271-2 specify internal arc tests to support the use of classifications (IAC) given to
metal-enclosed switchgear and related to provide safety to persons near the equipment if an
internal high power electric arc occur. The arc increases the pressure inside and this can even
open the doors, make a hole and eject hot gasses and materials which expose people outside
to a risk of life.
There are two design objectives. The first one to avoid that doors open - is easy to calculate.
To obtain the maximum pressures inside and to provide adequate means to avoid opening do
not require complex calculation tools. The second one is related to the fact that the metal-
enclosed switchgear need to have pressure relief window, usually in the top, in order to avoid
that the internal pressure goes beyond certain values. When the internal pressure exceeds a
certain value these windows open and the hot gasses flow outside. These gasses reflect on the
ceiling and external walls and can not go back to the body of a person outside which would be
injured. The distance the hot particles can reach can be estimated only by simulation
techniques or repeating trials in expensive high power tests.

The distance between the test and the real use:

The tests well represent what may occur in the case of an internal arc. The cotton pieces
which are installed to simulate the outside persons skin enable to clearly identify if the
equipment attend the standard requirements.
Example of points where the IEC standard is not clear enough to users and could be
improved:
One situation which may occur and it is not explicit is the case of a metal-enclosed switchgear
having capacitors filled with oil inside. In case of an internal arc involving also oil the
consequences would be much worst than not having oil. I do not know about tests already
done with metal-enclosed switchgear including oil capacitors but I know cases of such use.
Possibility of simulations in the design stage
The internal pressures are easy to determine considering the arc currents and the air or even
air plus oil volumes involved. It is possible to estimate where the hot particles and gasses will
arrive after the opening of the pressure relief window. With this simulation the manufacturer
can choose between different design alternatives and flows.
Possibilities to extrapolate the results of an already done laboratory test to other
untested equipment similar, but not equal
If the geometric dimensions of the untested metal-enclosed switchgear and distances to the
ceiling and walls are within a range of around minus or plus 30%, if compared to the tested
one, reasonable simulations can be done as an extrapolation.

3.2) TEMPERATURE RISE TEST AND HOT-SPOTS ON POWER

TRANSFORMERS ACCORDING TO IEC 60076

There are several IEC standards covering the different aspects of power transformers
specification and testing. These standards were published before 1993. In the last 20 years
there were technological changes in materials and design techniques. An example is that, in
the past, transformers tanks were much more resistant to the overpressures originated by arcs
inside the transformer tanks. It is interesting that these aspects are not covered by IEC
standards although transformers explosions and fires have hard consequences and are not a

4
rare fact. The Brazilian technical standard ABNT NBR 8222 / 2005 show guidelines for a
test related to the consequences of internal arc in the transformer with oil.

In the two next paragraphs there are two aspects of the standards IEC 60076 which
have a direct link with simulations and testing within the scope of this paper.

Temperatures and ageing in power transformers

IEC 60076-2:1993 Part 2 include requirements about temperature-rise limits and the
related type test. The test method for the temperature-rise test consists in measuring two
parameters. The first one is the average temperature rise of the windings above the oil (which
is above the external air temperature). This average temperature rise is calculated by knowing
the winding cold resistance without current and the hot one after passing the rated power and
temperatures stabilization. The second parameter is the temperature of the top of the oil
considered to be the highest inside the transformer. The ideal but not possible test procedure
would be to put thermocouples or something equivalent in the hottest points usually unknown
in advance.

Using the example of a 55C class transformer the fundament is that, for the limit condition of
an external air temperature of 40C, the maximum temperature in any point of the winding in
contact with the oil shall not exceed 40C + 55C + 10C = 105C. Above this 105C ageing
and oil decomposition start (table VI in IEC 60943).

Although not written in the standard, the 10C seems to be an estimative of the maximum
value above the average temperature of the windings that may be found in any part of the
winding. Possibly this value was obtained, more than 20 years ago, based on the existing
experience of the transformers manufacturers. Measuring systems and simulations in the 80s
where less accessible than nowadays.

Within the length of a transformer winding there are points with temperatures higher and
lower than the average value measured in the test. If we find in some part of this transformer
a 15C difference instead of 10C this could not be detected through the average temperature
measurement. It is not necessarily true that the top oil temperature is the highest anywhere in
the transformer.

If the fact described in the previous paragraph happens this means that a test approved
transformer may have inside, in winding parts in contact with oil, temperatures higher than
105C (oil decomposition). In the Annex B (informative) of IEC 60076-2: Part 2 it is
defined a hot-spot factor assumed to be from 1,1 in distribution transformers to 1,3 in
medium size power transformers or even higher in larger transformers. This seems to confirm
the possibilities of temperatures higher than 105C.

The question is: Are we sure that the 10C difference found more than 20 years ago is still
valid? If not this could be the reason for some recent failures in big transformers and reactors
with reasons not yet clarified? With simulations techniques it would be possible to advance
very fast in this respect.

Ability to withstand short circuit (thermal and dynamical effects)

5
IEC 60076-5:1976 Part 5 include requirements for power transformers to sustain
without damage the effects of external short circuits. It includes calculation
procedures used to demonstrate the thermal ability of a power transformer to
withstand overcurrents and both the special test and the theoretical evaluation
method used to demonstrate the ability to withstand the dynamic effects of the
short circuit forces . Experienced equipment designers know that, in the case of
the dynamical forces, such calculations are far from being considered exact and
can be checked only by testing. If we would test we would not need to calculate.

As the cost of transporting transformers to a high power laboratory is even bigger

than the expensive tests these calculations are the possible compromise and are
usually accepted. The calculation is useful only in the case of a future justice
demand. Even in this case nobody would assure that the calculation is right or
wrong because there is no recognized correct calculation.

The distance between the test and the real life:

The tests, in the two cases above, well represent the real life. We can not say that the method
to analyze the results of temperature rise tests in transformers is safe from the point of view of
premature ageing.
Example of points where the relevant IEC standard is not clear enough to users and
could be improved:
For the temperature-rise issue it is necessary to bring the information in the Annex B of IEC
60076-2: Part 2 to the main text and with objective requirements (and not informative).
Simulations can help to do this much faster.
Concerning to demonstrate the ability to withstand the dynamic effects of the
short circuit forces there is nothing to do and simulations will not bring new
facts.
Possibility of simulations in the design stage
For temperature-rise it is relatively easy to do the simulations but data to compare with is rare.
Possibilities to extrapolate the results of an already done laboratory test to other
untested equipment similar, but not equal
If the geometric dimensions of the untested transformer is within a range of around minus or
plus 40%, if compared to the tested one, reasonable simulations can be done as an
extrapolation.

4. SUGESTION FOR A STRATEGY, TO BE DISCUSSED WITHIN CIGRE, ON

HOW "TO ESTABLISH SOME RULES TO ENABLE SIMULATION TECHNIQUES
TO BE USED AS AN AUXILIARY TOOL FOR THE VERIFICATION OF SOME
TESTING REQUIREMENTS IN IEC STANDARDS"

As I am not sure about the format to present the proposal for CIGRE new work I will do it as
I was I used to do when I was coordinating the IEC Technical Committee TC 32 - Fuses.

The proposal for new work is to create a Working Group within IEC TC17 or TC32
involving both and also TC 14 (eventually other Technical Committees in a more advanced
stage) with the following task:

6
To establish some rules to enable simulation techniques to be used as an auxiliary tool
for the verification of some testing requirements in IEC standards"

It will be necessary to compare simulation results with known, reliable and published results.
At least in the initial stage of the work the Working Group will use as a reference, calculation
examples found in the Publication IEC/TR 60943 Guidance concerning the permissible
temperature rise for parts of electrical equipment, in particular for terminals.

The main tasks for the Working Group are:

Task 1) To define two test-cases of temperature-rise in electrical contacts (preferably using

existing examples in IEC60943). The published results are to be used as a reference for
comparison with the results obtained through simulations in existing tools. The test-cases
should have results confirmed by official test reports. The idea is that such test-cases may
serve in the future as the calibration when using a tool for simulation.

Task 2) To define one test-case of temperature-rise calculations in power transformers to be

compared with calculations made by simulations in existing tools. The case should have data
confirmed by official test reports. The objective is to find the temperatures of some known
points located in the winding and in contact with oil. The top oil temperature shall also be
calculated. The rated power of the transformer shall preferably be equal or higher than 10
MVA

Task 3) To list some additional test-cases which may be considered useful for future
verification of some testing requirements in IEC standards. Examples of these are the
calculation of overpressures inside metal-enclosed switchgear filled by air or inside
transformer tanks during an internal arc. Another example is to obtain the time-current
characteristics of expulsion and current limiting fuses

Task 4) To create a list of input data and output results to be used in simulations which may
permit comparison between simulations in different tools. An example of this is the input and
output data used in task 1 and 2 above.

4. CONCLUSIONS

Although simulation tools are still quite unknown to the average electrical engineer and
designer they are used for a long time by mechanical engineers. These are the tools which
made possible to design the wings of airplanes, equipment based on plasma, gas turbines and
fuel cells.

There is a clear space for their use in the electrical sector all over the World. Nevertheless, to
do this in an organized way requires some basic technical standardization.

A proposal to advance in this direction and to create a link between simulation tools, IEC
standards and testing laboratories is presented above to CIGRE.